1. Field of the Invention
The present invention relates to an image pickup device for picking up the image of an object and converting it into an image signal.
2. Related Background Art
As an image sensor for reading an original document, there is already known a contact multi-chip image sensor, composed of a light source, a short-focus imaging element array and plural line sensors. FIGS. 1 to 4 illustrate an example of such image pickup device. Referring to FIG. 1, at the upper end of a frame 200, there is provided a transparent glass plate 201 capable of contacting an original document. Illuminating light 212, emitted from a light source 210 provided in the frame 200, illuminates the surface of the original document contacting the upper face of the transparent glass plate 201. The light 213 reflected from the original is transmitted through an optical system 209 and an infrared cut-off filter 211 and is focused on a sensor array 1 provided on a substrate 19, corresponding to the optical system 209.
The above-mentioned optical system is composed, for example, of an array of imaging elements of a short focal length, such as Selfoc lens array (trade name of Japan Plate Glass Co., Ltd.). Also the above-mentioned light source is composed of a xenon tube or a cold cathode tube.
The sensor array 1 is constructed in the following manner. As shown in FIG. 2, the sensor array 1 is composed of a linear array of plural sensor chips 2, 2', 2", . . . on the substrate 19, and is covered by a protective film 206. The substrate 19 is supported by a bottom plate 205 engaging with the frame 200 as shown in FIG. 1, and is connected to a flexible circuit board 203 through a flexible cable 208. The flexible board 203 is provided thereon with a connector 202 for the input/output of power supply and control signals, and the connector 202 is mounted, as shown in FIG. 3, on the frame 200 with screws 207.
FIG. 4 shows the pixel arrangement in the vicinity of joints of the plural sensor chips 2, 2', 2", . . . arranged linearly on the substrate 19 constituting the sensor array 1. R, G and B pixels 3, 4, 5, respectively covered by R, G and B color filters are arranged in succession, with element isolation areas 6 therebetween, and these pixels release mutually independent signals. In the signal processing, the consecutive R, G and B pixels are regarded to constitute a picture element, and the output signals of the R, G, B pixels are regarded as R, G, B components of the picture element. Such arrangement enables image reading with a high resolution.
In the following there will be briefly explained the color reproduction of a color original, utilizing the contact multi-chip color image sensor of the above-explained configuration. FIG. 5 shows a CIE-xy chromaticity diagram, in which an area surrounded by a solid line, consisting of a spectrum line and a red-purple line, includes all the colors. FIG. 6 shows the spectral sensitivity distribution of the R, G and B sensor in the contact multi-chip color image sensor, FIG. 13 shows spectral emission characteristics of an Xe tube, FIG. 14 shows spectral transmission characteristics of an infrared cut-off filter and the chromaticity coordinates R, G, B of the sensors can be determined from FIGS. 6, 13 and 14. An area inside a broken-lined triangle, having the corner points at the chromaticity ordinates R, G, B, indicates the color reproduction area of the contact multi-chip color image sensor.
If R, G, B signals r, g, b are obtained in a picture element by reading an original document, the color of the original is represented by the coordinate of the center of gravity when the r, g, b values are respectively placed at the points R, G, B in FIG. 5. As an example, if the original document is the white reference, there stands a relation r=g=b, so that the color of the white reference is represented by the coordinate of the center of gravity of the triangle RGB, representing the color reproducible area of the contact multi-chip color image sensor. Similarly the colors of the picture elements present in the original can be determined and the color reproduction can be achieved with R, G, B colors.
In the following there will be explained the method of representing the resolution of an image. There are known various methods for representing the resolution, but, in the following, explained is a method utilizing MTF. By reading a grid pattern as shown in FIG. 7 with an image sensor, there is obtained an output signal as shown in FIG. 8. The modulation transfer function (MTF) can be obtained by substituting the maximum value i.sub.max and the minimum value i.sub.min of such output signal into the following equation: EQU MTF=(i.sub.max -i.sub.min)/(i.sub.max +i.sub.min).times.100 (%)
As the MTF of the sensor chip is generally sufficiently high, the MTF of the Selfoc lens array (SLA) used in the optical system is directly reflected in the MTF of the image sensor. FIG. 9 shows an example of the spatial frequency characteristics of such MTF.
For example, in case of an image sensor with a resolution of 400 DPI and with a conventional pixel arrangement as shown in FIG. 4, the pitch of the picture elements is equal to about 63.5 .mu.m, while the distance between the pixels R and G or G and B constituting a picture element is about 21.2 .mu.m, and that between the pixels R and B is about 42.3 .mu.m. With such distances among the pixels, the optical system 209 has a resolution of about 8 .mu.lp/mm between the picture elements, 24 lp/mm between the pixels R and G or G and B, and 12 lp/mm between the pixels R and B. According to FIG. 9, the MTF of the optical system is sufficiently low, about 0%, for a distance between the pixels R and G or G and B, but is about 30% at a distance between the pixels R and B, corresponding to a considerable resolving power which is no longer negligible in comparison with the MTF of 50% between the picture elements. Thus, if the optical system resolves the image light among the pixels constituting the picture element, there tend to be generated colored moiree fringes because of the difference in the optical positions, and there is provided a very unpleasant image particularly in case of a monochromatic image because of generation of false colors at the edge portion between black and white areas.
Such false color generation may be suppressed by reducing such colored moiree fringes for example through a signal processing so as to align the reading positions of the R, G, B pixels utilizing the outputs of the mutually adjacent pixels of a same color, but such method results in an increase in the circuit magnitude, because of the necessity for a memory for correcting the reading position, thus leading to an increased cost. Also the correction by such circuit alone cannot completely avoid the generation of such colored moiree fringes or false colors.
Also in case the sensor array is constituted by plural sensor chips, the spectral sensitivity characteristics may be different among such sensor chips, and the color space, constituting the basis of color reproduction, also becomes different among such sensor chips. For this reason, there may result an unnatural step difference in color, in the original image reading at the junction between the sensor chips. Particularly in case of an original which has a uniform color area over several sensor chips, there is reproduced a very unpleasant image with conspicuous streaks of color step difference.